OBJECTIVES: In a recent study in patients receiving warfarin, the initiation of statins that are cytochrome P450 3A4 inhibitors was associated with an increased risk of hospitalization for gastrointestinal bleeding in a Medicaid-insured population, while the initiation of pravastatin was not. The present research attempted to address some of the study limitations and assess whether similar findings are applicable in a more recent, commercially-insured U.S. population. METHODS: A retrospective matched-cohort design was used to compare baseline characteristics and bleeding rates (e.g. gastrointestinal and non-gastrointestinal) between statin and statin-free patients receiving warfarin concomitantly. Patients were matched on a 1:1 ratio to balance patient characteristics, and a cox proportional hazard model was used to control for confounding factors. The analyses were performed for a matched “stabilized” population, i.e., patients who had been on warfarin for ≥6 months, did not have prior bleeding events, and did not use another statin before the index date. For the matched “stabilized” populations, sub-group analyses were performed on patients with atrial fibrillation, patients with venous thrombosis, patients age ≥65 years, patients age ≥75 years, prevalent/persistent warfarin users, and by statin therapy. RESULTS: The method produced a small matched sample of 6306 (3.82%) out of 123,328 statin users and 41,734 statin-free patients. There were no statistically significant differences in bleeding rates between statin users and statin-free patients. Results were similar for the sub-group analyses. CONCLUSIONS: Using a claims database approach, high degree of heterogeneity between statin users and non-users was found, resulting in a low matching rate. This high degree of heterogeneity suggests that claims databases may be insufficient to detect/conclude as to differences in bleeding rates between statin versus statin-free patients on warfarin. Alternative methods and additional clinical information are needed to more accurately characterize bleeding rates in patients taking warfarin and a statin therapy concomitantly.